Group Title: evolutionary significance of certain behavioral, physiological, and morphological adaptations of the old-field mouse, Peromyscus polionotus
Title: The Evolutionary significance of certain behavioral, physiological, and morphological adaptations of the old-field mouse, Peromyscus polionotus
Full Citation
Permanent Link:
 Material Information
Title: The Evolutionary significance of certain behavioral, physiological, and morphological adaptations of the old-field mouse, Peromyscus polionotus
Physical Description: xiv, 187 leaves : illus. ; 28 cm.
Language: English
Creator: Smith, Michael Howard, 1938-
Publisher: University of Florida
Place of Publication: Gainesville, Fla
Publication Date: 1966
Copyright Date: 1966
Subject: Oldfield mouse   ( lcsh )
Microtus   ( lcsh )
Zoology thesis Ph. D
Dissertations, Academic -- Zoology -- UF
Genre: bibliography   ( marcgt )
non-fiction   ( marcgt )
Bibliography: Bibliography: leaves 171-185.
Additional Physical Form: Also available on World Wide Web
General Note: Manuscript copy.
General Note: Thesis - University of Florida.
General Note: Vita.
 Record Information
Bibliographic ID: UF00097880
Volume ID: VID00001
Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
Resource Identifier: alephbibnum - 000543141
oclc - 13108817
notis - ACW6847


This item has the following downloads:

PDF ( 7 MBs ) ( PDF )

Full Text

Peromyscus polionotus




June, 1966


During the period of this study, many recommendations

and constructive criticisms were received from various people;

to them, I express my appreciation.

I wish to thank Dr. D. Jameson because it was his en-

thusiasm that initially kindled my interest in zoology, and his

encouragement and support that prompted me to enter graduate

school. I would like to thank the members of my supervisory

committee, Drs. E. G. F. Sauer, B. K. McNab, H. M. Wallbrunn,

and B. N. Bunnell, for their patience, support, and suggestions;

Dr. J. Layne served on the committee for a short time and con-

tributed much to this work. I am especially indebted to my

chairman, Dr. E. G. F. Sauer; without him, the completion of

my work in the department would have been impossible.

Many other people contributed in various ways to my

work; A. Avenoso, D. Burnside, C. Collins, D. Crane, W. Criss,

M. Glenn, G. Gourley, R. McFarlane, C. Trost-, J. Wyrick, P.

Yingling, and my son, Michael W. Smith, assisted with the field

work and/or cared for the laboratory animals from time to time.

Paul Laessle, staff artist, supervised and helped in shaping the

illustrations. Doctor and Mrs. W. Dawson, Mr. R. Archbold, Mr.

and Mrs. R. Wyrick, and Mr. and Mrs. R. Lind kindly extended

their hospitality to me and made many aspects of the field work


more enjoyable and less expensive than they would have been

otherwise. C. Colson, R. McFarlane, and G. Phanuel critically

read parts of the manuscript. Doctors F. Nordlie and A. Laessle,

W. W. Bowen, and R. McFarlane discussed certain parts of the

paper at length with me. Mister Bowen also provided a copy of

a manuscript on subspeciation in Peromyscus polionotus that he

is preparing for publication. To all of the above, I express

my appreciation.

This investigation was supported in part by the American

Museum of Natural History, Sigma Xi, the Department of Zoology

at the University of Florida, a summer fellowship from the

National Science Foundation, and a Public Health Service pre-

doctoral fellowship from the National Institute of Mental

Health (1-Fl-MH-22, 122-01Al). I am grateful for this financial


Finally my wife, Irma Smith, deserves special acknowledg-

ment for the assistance she has given me. She helped with the

field work, the writing, typing and illustrating of this report,

care of the laboratory animals, and has provided moral and

financial support whenever needed. I sincerely appreciate her

many efforts.



ACKNOWLEDGMENTS . . . . . . . . . . ii

LIST OF TABLES . . . . . . . . . . viii

LIST OF ILLUSTRATIONS . . . . . . . . ... xi

INTRODUCTION . . . . . . . . ....... 1

Systematic relationships . . . . .... 1

ECOLOGICAL ASPECTS . . . . . . . . .. 5

Collecting Methods . . . . . . . . 5

Range of the old-field mouse . . . . . 5

Ocala National Forest. . . . . . . . 9

Habitat relationships. . . . . . . ... 11

Distribution of burrows soil characteristics .. 13

Description of an average burrow . . . .. 15

Local variation of burrows . . . . . . 17

Geographical variation in the depth of the burrow. 18

Depth of the nest cavity and seasonal temperature
fluctuations . . . .. . . . . 18

Ambient temperature and rainfall . . ... 22

Other species in burrows . . . . . .. 26


Laboratory colony . . . . . . . ... 32

Subspecific crosses. . . . . . . ... 33

Inheritance of white cheeks. . . . . ... 39




Assortative matings. . . . . . . .. 39

Calculation of inbreeding coefficient. . . .. .45


Torpidity in the field . . . . . ... 52

Torpor and availability of food. . . . .. .53

Concentration of respiratory gases and torpor. .54

Significance of torpor . . . . . . . 63


Results of field work. . . . . . . ... 65

Correlation between production of young, rainfall,
and ambient temperature. . . . . . ... 70

Effect of temperature on the male's reproductive
system . . . . . . . . ... . . 74

Minimum critical temperature . . . . .. 79

Breeding performance of Peromyscus polionotus
subgriseus in the laboratory . . . . .. 80

Onset of sexual activity . . . . . ... .80

Post-partum estrus and gestation period. ... .83

Synchronization of reproductive activity . . 85

Bruce effect . . . . .. ... ...... 88

Effect of litter sequence upon litter size . 89

Sex ratio at birth . . . . . . ... .90

Seasonal variation in litter size. . . . .. .92

Effect of light intensity and day length ... 93

Effect of food, temperature, and thyroid injections
upon production of young . . . . . ... .95

Geographical variation in litter size. . . .. .97





Dispersal of the Ocala population. . . . .

Predation in the Ocala National Forest . . .

Geographical variation in selection pressure .

Possible predators . . . . . . . .


General trends in pelage color . . . . .

Interpopulational variability of coat color in
north central Florida . . . . . . .

Interpopulational variation of external dimensions


IN CAPTIVITY . . . . .

Individual Behavior. .

Burrowing . .

Nest building .

Hoarding of food.

Escape behavior .

Social Behavior. . .

Communication .

Social units. .

Mating behavior

Parental care .

Discriminatory behavior . . . . . .











. . . . . . 132

. . . . . . 132

. . . . . . 134

. . . . . . 136

. . . . . . 138

. . . . . . 139

. . . . . . 139

. . . . . . 141

. . . . . . 141

. . . . . . 151



GENERAL DISCUSSION .................... .161

SUMMARY. .. ................ . . 168

LITERATURE CITED. ................. .. 171

BIOGRAPHICAL SKETCH. . . . . . . . . ... 186


Table Page

1. Collecting locality and habitat. . . . . . 6

2. Relative abundance of mice as it relates to the
physical and moisture retention data for some of
the common soil types found in Florida. Soil
data were taken from Stewart, et al. (1963) and
Bryan (1960) . . . . . . . .. . . 12

3. Animals found in burrows occupied (0) or un-
occupied (U) by old-field mice. The invertebrates
in unoccupied burrows were not identified. ... .27

4. List of food items commonly found in nest cavities 30

5. List and number of live animals placed into cages
with hungry old-field mice . . . . . ... .31

6. Analysis of crosses within and between different
subspecies of Peromyscus polionotus. The matings
included sexually active adults born and reared
in the laboratory unless otherwise noted. Wild
animals were paired in the laboratory in the same
combinations in which they were captured in the
field. Only data from pairs that remained to-
gether for at least three months during the fall,
winter or spring are listed below. . . . .. .34

7. Detailed statistical comparisons of the results
from the crosses presented in Table 6. Some of
the comparisons are interdependent but-the
associated increases in the P values were not
enough to alter the significance of the Chi
square values at the level indicated (.05 *
and .01 = **) . . . . . . . . . . 336

8. The number of times females were observed in or
out of their nest on the second, third, and fourth
nights following the birth of a litter. All ob-
servations were made with the aid of a red light
within two hours after the overhead lights went
off in the colony room. Chi square = 19.7 and P
was less than .01. . . . . . . . ... 38


LIST OF TABLES (Continued)

Table Page

9. Inheritance of white cheeks in Peromyscus
polionotus. The expected values for the
statistical analysis were calculated by
assuming that the inheritance of the trait
was due to a single gene with two alleles,
Wc and wc. . . . . . . . . .. . . 42

10. The frequence of mice with white or brown cheeks
and the observed and expected number of paired
mice with these phenotypes in the Ocala National
Forest and at Archbold's Biological Station in
Highlands County, Florida. Expected values were
calculated from the observed phenotypic frequencies
and Yate's correction for continuity was used in
the Chi square analysis of the data for the popula-
tion from Highlands County to compensate for the
small sample size of some cells. . . . . ... 44

11. The number of times brown-cheeked females mated
with brown-cheeked siblings, white-cheeked non-
siblings, white-cheeked siblings and brown-
cheeked non-siblings. The choice always involved
either the first two possibilities or the later
two and no combinations thereof. Expected values
were calculated for the Chi square analysis by
assuming the null hypothesis. They had to be
adjusted for the choice involving the heterozygous
white-cheeked sibling because a certain percentage
of the litters sired by this animal would contain
all brown-cheeked offspring, and thus, would be
counted for the brown-cheeked non-sibling. ... .46

12. Variations in the females' reproductive character-
istics. Pregnant or perforate females were con-
sidered to be reproductively active. . . . 66

13. Monthly variation in reproductive characteristics
of adult males . . . . . . . . . 68

14. Breeding histories of five typical laboratory-
reared females. The size of each litter is
indicated in parentheses after the date of birth .84

15. Effects of four different treatments upon the
reproductive performance of laboratory-reared
mice. The group injected with thyroxine had
its own control group because this experiment
was started after the others had been completed. . 96

LIST OF TABLES (Continued)

Table Page

16. Social units found in excavated burrows of
Peromyscus polionotus subgriseus . . . ... 142

17. Mating data for laboratory-reared Peromyscus
polionotus subgriseus. . . . . . . .. .146

18. The nightly distribution of mice of Peromyscus
polionotus and P. gossypinus in a cage with four
compartments .. . . . . . . . . .. 156

19. The nightly distribution of mice of the two sub-
species, Peromyscus polionotus subgriseus and P.
E. phasma, in a cage with four compartments. ... .157

20. The nightly distribution of mice in a cage with
four compartments. The tests involved two
species, Peromyscus polionotus and P. gossypinus,
or two subspecies, P. p. subgriseus and P. p.
phasma . . . . . . . . ... . . 158


Figure Page

1. The approximate range of the old-field mouse,
Peromyscus polionotus. . . . . . . . 2

2. Relative abundance of burrows on different parts
of old sand dunes in the Ocala National Forest . 14

3. Lateral view of the burrow of the old-field mouse,
Peromyscus polionotus. . . . . . . ... 16

4. The depth of the bottom of the nest cavity at
various localities. Mean values are given plus
or minus the standard error. The number of
burrows at each locality is given in parentheses 19

5. Temperature at various depths and at two dif-
ferent times of the year. These months were
chosen because they show the maximum variation
in temperature recorded at the surface of the
ground and the typical shape of the winter and
summer curves. . . . . . . . . .. 21

6. The range and mean temperature of the air in the
nest cavity and the mean temperatures of the soil
at the depth of the nest cavity and the air above
ground in the Ocala National Forest. . . .. 23

7. Air temperature (above ground) and rainfall
deviations from the expected values as calculated
by the United States Weather Bureau for Ocala,
Florida. . . . . . . . . . .. . 24

8. The regression of the amount of rainfall on the
temperature of the air above ground each month
in the Ocala National Forest . . . . .. 25

9. Geographical variation in the percentage of mice with
white cheeks. Only samples of 15 or more mice are
included . . . . . . . . . .. . 40

10. Metabolism apparatus . . . . . . . 56


Figure Page

11. Metabolic rates of active (open circles) and
inactive (solid circles) Peromyscus polionotus
at various oxygen concentrations. All of the
animals survived down to 2 percent oxygen at
which time the experiment was terminated.
Values higher than those given above were also
recorded . . . . . . . . .. . . 58

12. Metabolic rates of Peromyscus leucopus at dif-
ferent oxygen concentrations. Values higher
than those given above were also recorded. None
of the mice survived below 4 percent oxygen. ... .59

13. A. Average monthly rainfall at Ocala, Florida.
B. Percentage of mature females that were re-
productively active and average litter size per
month. The monthly sample sizes are given across
the top of the figure. The criterion of maturity
was the lack of the juvenile pelage or the pres-
ence of embryos. . . . . . . . . ... 71

14. The size of the seminal vesicles and testes, and
weight of the testes and body of males housed with
other males at four different ambient temperatures.
There were nine mice at each temperature. The mean
is indicated by the horizontal line through the
dark rectangle which represents two standard errors.
The range is shown by the other horizontal lines,
and the open rectangles indicate one standard de-
viation on either side of the mean . . . ... 76

15. Breeding performance of laboratory-reared
Peronyscus polionotus subgriseus maintained
at 240 1 l0C . . . . . . . .. . . 81

16. The number of litters born on each day during
December, 1964. The numbers in each block repre-
sents the shelf on which the birth occurred.
There were five cages on each of 12 shelves. ... .86

17. Average size of sequential litters born to
laboratory-reared Peromyscus polionotus
subgriseus. Only females which had three or
more litters were included. The number of
litters is given above each point. . . . .. .91


Figure Page

18. Dispersal of juvenile males and females
(Peromyscus polionotus subgriseus) . . . .. 100

19. The ratio of the number of juveniles to adults
and of adult males to adult females in each
monthly sample from the Ocala National Forest. . 105

20. Geographical variation in coat color of Peromyscus
polionotus. The numbers are on an ordinal scale
with one being the lightest and eight and darkest. 117

21. Variation in the size of the ear in 17 populations
representing six subspecies (separated by dashed
horizontal lines and indicated by letters inside
the open rectangles; L = Peromyscus polionotus
lucubrans; P = P. p. polionotus; PH = P. p. phasma;
S = P. p. subgriseus; R = P. p. rhoadsi; N = P. p.
niveiventris). The number of each population as
given on the map corresponds to the number at the
right of the histogram representing it. The
vertical line denotes the mean, the horizontal
line the range, the closed and open rectangles
one standard error and one standard deviation on
either side of the mean. . . . . . . ... 122

22. Variation in the size of the hind foot in 17
populations representing six subspecies
(separated by dashed horizontal lines and
indicated by the letters inside the open
rectangles; L = Peromyscus polionotus
lucubrans; P = P. p. polionotus; PH = P. p.
phasma; S = P. p. subgriseus; R = P. p.
rhoadsi; N = P. p. niveiventris). The number
of each population as given on the map cor-
responds to the number at the right of the
histogram representing it. The vertical line
denotes the mean, the horizontal line the range,
the closed and open rectangles one standard
error and one standard deviation on either side
of the mean. . . . . . . . . ... 124

23. Variation in the size of the body in 17
populations representing six subspecies
(separated by dashed horizontal lines and
indicated by the letters inside the open
rectangles; L = Peromyscus polionotus
lucubrans; P = P. polionotus; PH = P. p.
phasma; S = P. p. subgriseus; R = P. p.
rhoadsi; N = P. p. niveiventris). The number



Figure Page

of each population as given on the map cor-
responds to the number at the right of the
histogram representing it. The vertical line
denotes the mean, the horizontal line the
range, the closed and open rectangles one
standard error and one standard deviation on
either side of the mean. . . . . . . ... 126

24. Variation in the length of the tail in 17
populations representing six subspecies
(separated by dashed horizontal lines and
indicated by the letters inside the open
rectangles; L = Peromyscus polionotus
lucubrans; P = P. . polionotus; PH = P. p.
phasma; S = P. p. subgriseus; R = P. P.
rhoadsi; N = P. g. niveiventris). The number
of each population as given on the map cor-
responds to the number at the right of the
histogram representing it. The vertical line
denotes the mean, the horizontal line the range,
the closed and open rectangles one standard
error and one standard deviation on either
side of the mean . . . . . . . . .. 128

25. Seasonal variation in the percentage of nests
of Inboratory-reared mice wi thout roofs on
the second day after the birth of a new litter . 135

26. Percentage of nest without roofs during the
month following the birth of a litter to
laboratory-reared parents. . . . . . .. .137

27. Sequence of behavior patterns during mating.
Solid lines indicate the usual sequence of
behavior leading up to and following the first
few copulations. Broken lines indicate alter-
nate sequences used at various times during
mating . . . . . . . . . . . 144

28. Cage used for measuring discriminatory
behavior . . . . . . . . .. . . 154


The old-field mouse, Peromyscus polionotus, shows more

variation in morphology within a limited geographical area than

any other species in the genus (Blair and Howard, 1944; Hayne,

1950; Blair, 1951; Schwartz, 1954). The degree of divergence

within a species is limited by the amount of genetic exchange

between populations and by the nature and intensity of the selec-

tive pressures they are being subjected to (Dobzhansky, 1951;

Mayr, 1963). This study represents an attempt to gain an under-

standing of some of the effects and interactions of variation,

selection, and isolation as they relate to the intraspecific

evolution of the old-field mouse.

Systematic relationships

Peromyscus polionotus is a small, semi-fossorial, nocturnal

rodent of the subgenus Peromyscus and the maniculatus species group

(Osgood, 1909; Blair, 1950; Hall and Kelson, 1959). It is endemic

to the southeastern United States (Fig. 1) and may have been iso-

lated from its parental stock, P. maniculatus, since the period

of Pleistocene glaciations (Blair, 1950). There are two other

species of Peromyscus in Florida, P. floridanus of the subgenus

Podomys and P. gossypinus of the subgenus Peromyscus. Ochrotomys

nuttalli is excluded because it is no longer considered a Pero-

myscus (Hooper and Musser, 1964).


Fig. 1. The approximate range of the old-field mouse,
Peromyscus polionotus, in the southeastern United States.


/ .:.* ''. : .".. ..: :""" v'<. "- -.

.. : .. -.. .. .

I ES, N \


.,, ..

O .-S-0
\ ...

00 .: .... .: .


There are no known hybrids between species of different

subgenera of Peromyscus. Peromyscus polionotus is able to hy-

bridize with P. maniculatus in the laboratory, but it does not

hybridize with P. gossypinus or P. leucopus (Dice, 1933; Watson,

1942; Dawson, 1965).. Possible hybridization occurs in the field

between P. maniculatus and P. leucopus (Waters, 1963) and between

P. leucopus and P. gossypinus (Howell, 1921; Dice, 1940; McCarley,

1954; Golley, 1962). It appears that P. leucopus, P. gossypinus,

P. maniculatus, and P. polionotus are closely related. Even if

this were not the case, it is evident that P. polionotus is more

closely related to P. gossypinus than to P. floridanus (Hooper and

Musser, 1964). Despite the close relationship of P. polionotus to

several other species, it can be easily identified as the smallest

species of Peromyscus in the United States (Blair, et al., 1957,

p. 708).


Collecting methods

Many species of Peromyscus spend the daylight hours in

an underground nest. However, P. polionotus is exceptional

among the semi-fossorial members of this genus in that it digs

its own burrow and does not depend upon the chance occurrence

of suitable shelters. The burrow entrances are conspicuous,

thus, facilitating collection of these animals in the field

(Sumner and Karol, 1929; Hayne, 1936; Smith, 1939; Rand and

Host, 1942; Laffoday, 1957; Smith and Criss, in press).

Range of the old-field mouse

A revised range map was drawn up using the field data,

as well as pertinent references (Chapman, 1893; Bangs, 1898a,

1898b; Osgood, 1909; Howell, 1921; Sumner, 1926; Sumner and

Karol, 1929; Dice, 1934, Hayne, 1936; Coleman, 1939, 1948;

Rand and Host, 1942; Moore, 1946; Ivey, 1949; Schwartz, 1954;

Hall and Kelson, 1959; Golley, 1962), and data from specimens

in the Florida State Museum. Collections were made in Florida,

Georgia, and South Carolina in places where it was reasonable

to expect that a sample of 20 adult animals could be taken

(Table 1). Otherwise, the geographical location of the burrows

was noted. The revised map differs from that of Hall and Kel-

son (1959) in several respects (Fig. 1). The mice occur as far


O r0

o 0 Z

4 > r m
0 0H C

! > c ) d 0
CnC C dC)
3 ^ (
)--' CCooi i






0 "

0 0

0 0




C cae

CC 0 H

-0 >0 o CC 1)
*H H Cn o

U O OJ C 8 H C
I- C C)d i s i' t
ro (l al i 3 *HO
^ > ^ X tii O U 0


0 C)

v H C

o C a
'N C)l N

'C C)
C) .
'c' c' 0
~ 0 C)

*o 'o 4o po 0 4
CC aC C C) a 0H

-1 0 3 0 0
H O C O O H O H0 a
) In C C ) aw 0

- H Co a -
'C o O I0 r' O)
HO 0 O 4 C )
0C C C o HOC)



t r


S *H 1 0 "

o a (5 0
1 -H >
0 r 0 4 C
0 o as a)
o 0 4- r-
o O


-4 U



iUl '0

H~ 0
X Hl
*r-4 *H


o 0

0 r 4
0 C0 -

0) Cn
^ (O
F0. 0

> *H *H
(( H

. H 4
0 u 0 1
2 U U en

0 O
. 0
o "

) UC )
'0 'U
o 3 0

U 1

a o )

co a) 0



8 0

0 a0

o a ..
= '= 0
0 Q a 0

o r c o
Sl U N


a) a

0 > 0

L. < S






' 0

V 0
i 4
1 -1


H *H

oH U


0 '0



c n~

*O r



C -





U) 0

0 0

H *H
> U 0

C "o H





U) H0

f ~co
0 n
w H

south as Highlands County in the central part of Florida, and

they were not found in the Cedar Key area, nor around the Oke-

fenokee Swamp, nor east of the Saint Johns River except on the

sand dunes along the beach. The locality south of Camden, South

Carolina, is a new range extension (Table 1, Fig. 1).

Ocala National Forest

The majority of the field work was done in the 146,497

hectare (362,000 acre) Ocala National Forest located on the Florida

peninsula between Ocala and Daytona Beach. The area is bounded on

north and west by the Oklawaha River, on the east by Lake George

and the Saint Johns River, and by state highway 42 on the south.

Parts of Marion, Putnam, and Lake County make up the forest.

Approximately 184 km (115 miles) of paved roads and about four

times as many maintained dirt roads lead through the forest.

The grass on the road shoulders is mowed at regular intervals.

About 99 percent of the specimens came from burrows on the road

shoulders where they were concentrated and conveniently located.

To minimize the effects of sampling, mice were never taken

twice from the same stretch of road within any six-month period.

Some active burrows were not touched to insure repopulation of the

area. Along one five-mile section of highway 316 between Eureka

and Salt Springs, all the burrows were dug out and the occupants

captured. New burrows were excavated on subsequent field trips.

The approximate age, sex and reproductive status were recorded for

each animal. Females that were perforate, pregnant, or lactating

and males that had their testes in a scrotal position or sperm

in the epididymes were considered sexually active. Sexual

inactivity was indicated by the absence of these conditions (also

see Smith and Criss, in press).

Twelve to 36 occupied burrows of the old-field mouse were

dug up in the Ocala National Forest plus or minus ten days from

the first of each of the following months: December, 1962,

April, July, August, September, October, November, and December,

1963, January, February, March, April, May, June, September,

October, November, and December, 1964, January, February, March,

April, July, August, September, 1965, and on May 15, 1965. The

number of occupants, approximate age, sex and reproductive status

of each occupant, presence or absence of a nest, or uneaten food,

and other species of vertebrates and invertebrates in occupied

burrows were noted. The temperature of the soil at the depth

of the nest cavity was recorded for at least five occupied burrows

each month. From July, 1963, to June, 1964, the temperature and

occasionally the percent oxygen of the burrow air in the nest

cavities were recorded. Oxygen concentration was measured with

a Beckman oxygen electrode. Air temperatures and the percent

oxygen in the nest cavity were obtained by taping the probes

onto a two-foot rod and inserting it down the entrance tube into

the nest cavity immediately after the sand plug was removed.

The entrance tube was again closed with sand, and the readings

were taken within three minutes. The dimensions of the five

burrows were recorded during each of the 12 months. The angles

between the horizontal plane and the escape tube as it left the

nest cavity or the entrance tube just behind the sand plug were

measured with a protractor and a level.

The mice sacrificed in the field were placed in a plastic

bag and kept on ice. The width and length of the testes of each

adult male were measured with a caliper; the epididymes were

microscopically checked for sperm, and the length of the seminal

vesicles were measured after the method of Jameson (1950). Fe-

males were checked for the presence of embryos.

Habitat relationships

Mice were collected in a variety of.habitats primarily

in north central Florida (Table 1). Monk (1965) divides the

vegetation of this area into swamp forests, mixed deciduous and

evergreen forests not periodically flooded, and pine forests.

The mice were never found in swamp forests, nor in forests pre-

dominated by evergreen trees. The pine forests, which are fire

sub-climaxes, are subdivided into sandhill, sand pine scrub, and

pine flatwoods (Laessle, 1958a; Edmisten, 1963; Monk, 1965).

The pine flatwoods frequently occur on poorly drained soils; sand

pine scrub and sandhill are confined to well-drained upland soils.

The mice were found in old-fields, cleared farmland, sand dunes

along the beach, sand pine scrub, sandhill, and pine flatwoods

(mostly long leaf pine) in decreasing abundance in the order


Abundance is correlated with soil type, amount of soil

drainage (Table 2), type and amount of vegetation. All of the

habitats occupied by this species are characterized by sparse

vegetation and relatively well-drained or recently plowed soils,

except pine flatwoods. In this habitat, burrows were rarely

found, and the mice were probably transients; no permanent popu-


0 0H


1 0

n e





,a 0
( C
* d
> o

CO *i C

IH 0




+ -


U )



0 c

4 *-


Q )

H ) C

C)0)U Q
I-' C )



C n o

N 10 0 H
C" C* H* ' '

\o ^

'0 d

-l 0C


H 00 0


n co o'
*^ ii Hi O>

lations were found in the pine flatwoods. An inverse relation-

ship between the density of animals and the amount of ground cover

may exist (Rand and Host, 1942).

Distribution of burrows and soil characteristics

Certain soil characteristics appeared to be important in

limiting the distribution of mice. In relatively undisturbed

habitats, the mice occurred primarily on fine sand including the

well-drained, acid Saint Lucie, Lakewood, Lakeland, and Kershaw

soils which have a low clay content, low moisture retention

capacity, and high hydraulic conductivity (Stewart, et al.,

1963; Table 2). Mice never occurred in high densities on poorly

drained soils. In addition, they never constructed burrows in

hard soils where digging was difficult, nor in areas where the

hardpan was close to the surface of the ground.

Burrows were built on the sloping banks of sand dunes

on the beaches. In the Ocala National Forest, they were usually

found on the slopes of sand hills, probably remnants of old

sand dunes (Laessle, 1958b; Fig. 3). Occasionally, they were

located on the road shoulders at the top of these sand dunes

but rarely in the low ground between the dunes.

Deposits of sorted sands have been laid down in several

ways in Florida (Laessle, 1958b). Wind was an important agent

along the beach dunes. The action of water was important along

the flood plains of large rivers, the shore lines of lakes and

small islands, and submerged offshore bars. All areas above the

current water level were at one time part of the Florida shore-

line. As the water level fell during glacial periods, numerous

; * o ."

. ... .. . .o


*. 0

\', '. -' ."* -o
. . 0



:.'. ." o '
.. ..
: 44

-.." .*

:.' .' . .. .*. X)
--N 'H

'.' .. . 0
*.:.:. : '. :... o
** ~. 0I

* 4
.: . (C

deposits of fine textured sand were gradually exposed. Their

continuity was later destroyed by erosion (Alt and Brooks, 1965).

These deposits and their associated vegetation are frequently

widely spaced with the intervening habitat unsuitable for the old-

field mouse. These interrupted sand deposits are ecological

islands for this species.

Description of an average burrow

Over 1,500 occupied burrows were excavated. An average

burrow consists of an entrance tube, nest cavity, and escape

tube (Fig. 3). The three parts are aligned almost in a single

vertical plane with the nest cavity in the center. The entrance

tube, closed with a sand plug of 12.8 1.3 cm (range:2.5 to

61 cm; means are given plus or minus one standard error), levels

off horizontally before reaching the nest cavity. The escape

tube ends a few centimeters below the surface of the ground

under or near a clump of grass. The angle of descent of the

entrance tube at a point just behind the sand plug is 38.20 +

1.80 (range:240 to 480) from the horizontal. The escape tube's

ascending angle as it leaves the nest cavity is 64.30 2.1

(range:440 to 790). There is a significant positive correlation

between these two angles (r = .86, df = 59, and P less than .01;

statistical methods were adapted from Steel and Torrie, 1960).

Mice tend to dig a shallow or a steep escape tube depending upon

whether the entrance tube was shallow or steep, respectively.

The volume of each part of the burrow was calculated by

assuming it was a cylinder, which shape they did approximate.


.c.. .. .. .

. .

. a . .
: '.

. :*
," " .''"


.. ". .
. .. '.


:::i:" "" '

'. ;.'

... ..: : . .

t ".b

':. ".:'

The formula was V =n D2L/4, where V is the volume, D the diameter

of the tube, and L its length. The escape tube (102.1 + 6.8 cm)

is almost twice as long as the entrance tube (68.1 3.7 cm) and

about ten times the length of the nest cavity (10.7 .4 cm).

The sum of the three volumes minus the volume of the sand plug

is equal to the volume of gas in the burrow which averaged

4,046 + 231.8 cc.

Local variation of burrows

The slope of the ground and the presence of underground

obstacles were the primary physical factors which altered the

profile of the burrows. Abrupt changes in the course of the

burrows were correlated with changes in the slope of the ground.

They normally went toward the high ground in the same direction

as the slope at the surface of the ground. This variation was

common along road shoulders where the direction of slope changed

900 within a few centimeters. Burrows also curved around ob-

stacles, such as roots, and then continued in their previous


Many times adjacent burrows deviated from the average

in the same way, e.g., two escape tubes or a very large nest

cavity. On three different occasions, several fresh burrows

were excavated in a limited area and only one female, molting

from the juvenile to the subadult pelage, was captured. No

other burrows were found in the immediate vicinity, and each

of the excavated burrows had two escape tubes. Once a solitary

adult male was captured in one of five fresh burrows in which

each of the escape tubes curved back toward the entrance. On

another occasion, an adult female with a litter was found in one

of three fresh burrows which all had the escape tube coming off

the entrance tube at its junction with the nest cavity. It

seemed likely that each captured mouse had constructed all of

the burrows in its area, and that each one consistently dis-

played its own individual peculiarity.

Geographical variation in the depth of the burrow

Only quantitative differences were found between the bur-

rows at the different localities. The depth of the nest cavity

is discussed here; it was measured from the bottom of the nest

cavity to the closest surface of the ground (Fig. 4). At certain

places, e.g., Manatee Springs, Florida, the depth of the nest cavity

was limited by the depth of the hardpan, but at most localities,

e.g., south of Camden, South Carolina, the hardpan was much

deeper than the deepest nest cavity. The shallowness of these

burrows may be explained by variation in some other soil character-

istics, possibly clay content, which may influence the rate of

diffusion of respiratory gases through the soil (Penman, 1940).

Depth of the nest cavity and seasonal temperature fluctuations

With the seasonal variation in soil temperature greater

at shallower depths than at deeper ones, the nest cavities were

located just deep enough to take advantage of the maximum damping

effect of the soil on these seasonal ambient temperature fluc-

tuations (Fig. 5). The difference between the seasonal tempera-

ture fluctuations at the average depth of the nest cavity and at

a point twice this deep was less than 10C. Diurnal fluctuations

are also negligible at this depth (Vorhies, 1945; Schmidt-Nielsen

Fig. 4. The distance from the bottom of the nest cavity to the
surface of the ground at various localities. Mean values are
given plus or minus the standard error. The number of burrows
at each locality is given in parentheses.

33.0 +2.7 (14)0

41.4 2.8 (8)0

46.2 10.9 (7)0

47.3 11.4 (5)o

31 + .\
40.6 2.1 .(13)o 4 3 1.9 (17')
44.9 -1.4 (27)0 O-.----. -- '
.. 45.8 -1.2 (31)0 .
63.5. 4.0 (150n
-30 50.3 .2 (18)0 +
+ -,458 1.8 (8)
5.4 8 (17) 41. 5 2.1 (14)
47.4 -1.8 (12)0 0 054.5 2.4 (15)
35.9 -1.7 (20)0 4.3N (60
48.4 -1.3 (41) 0'54"31. (60)
-9 MILES 49.8 3.0 (27)0 058.6 -2.6 (2/
0 1 00 47.6 1.0 (6)0

56.3 -2.2 (19)0 (8)
-28 54.8 2.2 (13)

59.1 .9 (43)0
27 \

-26 87 86 85 84 83 82 81
I I l I l

-35 ... ..... .

I I I I .. .

t i

(o0) 3J1nv1iV3dJN3i

U) )



0 '

C- 0

o S4
0 1

o ,

W 0

0 0:
m U.) H
- 4:2

Lo o
-J O

0 4"9 0

00 9

+ 01 41
0, fl c

o c o

0 0
C) a

*d C 0 C)

-4 ai

i 4-
4-> g *

and Schmidt-Nielsen, 1950; Petter, 1952; Prakash, et al., 1965).

Digging a deeper burrow would not provide the mice with a more

constant thermal environment, but this would require them to

expend more energy and increase the layer of soil through which

the respiratory gases must diffuse.

Ambient temperature and rainfall

The soil temperature at the depth of the nest cavity

showed significant positive correlations with the mean monthly

air temperature in the nest cavity and above ground at Ocala,

Florida (r = .999, df = 11, and P less than .01 and r = .902,

df = 28, and P less than .01, respectively; Anonymous, 1962,

1963, 1964, 1965; Fig. 6). The air in the nest cavity ranged

from 120 to 340C, and it averaged l.O0C higher than the soil

temperature at this same depth in both occupied and unoccupied

burrows. This indicates that the temperature difference was not

due to heat produced by the tice, nor by decay of organic matter.

It was approximately the same in the summer when little or no

organic matter was found in most of the burrows.

During this study, the air temperature and rainfall at

Ocala averaged .30C lower and .3 cm greater, respectively, than

expected in an average year (Fig. 7). A significant positive

correlation existed between the amount of rainfall (Y) and the

mean air temperature (X) for each month (r = .425, df = 30, and

P less than .05; Fig. 8); the equation of the regression line

was Y =.78X-3.6. The magnitude of the monthly deviations in rain-

fall and air temperature were not correlated (r = .203, df = 30,

P greater than .05), but there was a significant tendency

0 -




> ri


C) 0
4-) 4

c 0






3 C


0 0

C 0Z
- *i'
CO 04'


SCO ID IC 0 i O i -
+ + + + + +
I s *


r i

t I I




0 0

e O
1 )




L.- ri
-Q 1'-
z CC

I- *rla



-- oC

to A ro N 0 N nr
+ + + +nt NI NO I

(Do) 3fnlV83dLN3 NI NOIIVIA30






- I I I I I I

t 0 ao 0 L
r( Vf) (l ) N l -

(N~o) 1-1VA NI IV6

0 n 0



O -


0 0




0 0

'O 0


to m

01 0

for the deviations to occur in opposite directions (Chi square =

3.98, df = 1, and P less than .05; Fig. 7). Seasonal increase

in rainfall was associated with higher temperatures but rainfall

greater than the monthly average was associated with air tempera-

tures lower than the monthly average, and vice versa.

Other species in burrows

Invertebrates found within occupied burrows or in the

entrance tube outside the sand plug, and vertebrates in unoc-

cupied burrows were identified (Table 3). The only animal that

was regularly found in burrows with old-field mice was the camel

cricket. It builds small lateral tunnels off the sides of the

escape tube. The crickets have to avoid the mice since they are

readily eaten by them. Two blind cricket-locusts were also found

with the mice. Both of these species have also been reported from

pocket-gopher burrows (Hubbell, 1936, 1940; Hubbell and Goff,

1939). The commensal relationship between Typhloceuthophilus

floridanus and Geomys was thought to be obligatory (Hubbell,

1940). This cannot be true since the collecting sites of the

two specimens were both far from the nearest Geomys burrow.

A number of vertebrate species inhabited abandoned bur-

rows. Coachwhips were the most frequently encountered species

of snake. All of the snakes found in burrows, except the common

hognose and the eastern diamondback rattlesnake, were shedding.

Two species of mice, Mus musculus and P. gossypinus, used the

unoccupied burrows. The house mice occurred in the burrows only

on farmlands. The same was true for the least shrew and the east-

ern mole. The P. gossypinus were found in one burrow in the sand

C0 > > 0C ,OH

"C .. 0 -

H 0H

.o 4- (A)

0 -
H1 1 41 |

*C) U V

|o t i 1 .!
, C) a) I 0 0V

0o H W ( t> o W
0 4 l 4 4 o I H H 0 f

H 0 a) a) a) C)

0-H 0 C C o X o
H -H 3 C -H
, a H H -

(CC () 0 i r-o H C

.00 c 0 CC CC H CC

0 0.

CC .4 (C

+ j 0 0 a-C 01
r C +CC 0 CC

.0 U H *H C

CC i Hx C a) C

r-( CCCU) ro H 00 H

.H 0 0 4 (6

H ca C () 0 V) C

*H *H *H *H *d *H
> > > > > >

c) V) ) ) V)

x x x x x x x x

-4 M -4- H- H H -4 r-



'C c


O4 H
o H
H *i
U(- ia
O '
*r4 *H
H 'C
o HI

0 Tl

H .0
o 1

C) c
< 0

H 'C
m c
m ui


* 0


0 I

a4 V
o 0

x x x x x x

-H c 0 N a) N

> H

H H 0
- C)

c 0 '

S 'C 0
. 0

Cd 'C
0 U

C) C)

4 4
4 -H

o o

a( ai
c c

pine scrub habitat within the Ocala National Forest.


Peromyscus poliono'us is omnivorous like the other species

of Peromyscus that have been investigated (Jameson, 1952; Connell,

1959; Williams, 1959). No attempt was made to identify all of

the various foods eaten by the mice but the more conspicuous

items found in the nest were identified (Table 4). Many different

types of organisms were offered to hungry mice in the laboratory

(Table 5). They ate all of them except the toads. Live animals,

such as cockroaches, were frequently fought over and seemed to

be preferred over most seeds. In the field, more fragments of

nuts and seeds were found in the nests than animal parts. Most

of the plant material consisted of empty shells which do not

decay as fast as other materials.

Table 4. List of food items commonly found in nest cavities

Plant Matter

Acorns from Myrtle Oak (Quercus myrtifolia)

Acorns from Chapmans Oak (Q. chapman)

Acorns from Sand Live Oak (Q. virginiana)

Acorns from Turkey Oak (Q. laevis)

Seeds from Partridge Pea (Cassia fascioulata)

Animal Matter*

Beetles (several types); Carabidae and others unidentified

Beetle Larvae in Acorns; Curculioninae (Balaninus sp.)

Lepidopterans; Danaidae, Pieridae and others unidentified

Eastern Lubber Grasshopper (Romalea microptera)

Grasshoppers (several other types); Locustidae

Dragonflies (several types); Aeschnidae

*the animals were all partly eaten so complete identification was
usually impossible.

Table 5. List and number of live animals placed into cages with
hungry old-field mice

Species Number Fate


Black Widow (Latrodectus mactans) 4 eaten

Carolina Wolf Spider (Lycosa carolinensis) 3 "

American Cockroach (Periplaneta americana) numerous "

German Cockroach (Blatella germanica) "

Camel Cricket (Ceuthophilus latibuli) "

Blind Cricket-locust (Typhloceuthophilus
floridanus) 1 "

Common Field Cricket (Gryllus assimilis) 8 "

Eastern Lubber Grasshopper (Romalea microptera) 3 "


Southern Toad (Bufo terrestris) 6 not

Oak Toad (Bufo quercicus) 5 "

Six Lined Racerunner (Cnemidophorus sexlineatus) 3 eaten

Scrub Lizard (Sceloporus woodii) 2 "

Cotton Mice
(1 day old) (Peromyscus gossypinus) 4

Florida Deer Mice
(1 day old) (P. floridanus) 3 "

Old-field Mice
(1 day old) (P. polionotus) 4 "


The degree of isolation between various species and

subspecies of Peromyscus is highly variable. In some areas,

genetic exchange is minimal between different subspecies, while

in others, hybridization occurs frequently between recognized

species (Howell, 1921; Dice, 1933 and 1940; Dice and Liebe, 1937;

Watson, 1942; Blair, 1944; Liu, 1953a and 1953b; Harris, 1954;

McCarley, 1954; Barbehenn and New, 1957; Sheppe, 1961; Golley,

1962; Waters, 1963; Dawson, 1965). The results of this inter-

breeding are important in determining the evolutionary course

of the factors influencing genetic exchange at the population

and subspecific levels in P. polionotus.

Laboratory colony

During 1962 and 1963, 64 pairs of mice, as they were

captured in the field, were used in the laboratory for breeding

stock. Occasionally, one escaped or died and had to be replaced;

additional pairs were added to the colony. Four Florida sub-

species were represented in the laboratory colony, P. polionotus

subgriseus from Ocala National Forest, P. p. phasma from

Anastasia Island, Saint Johns County, P. p. leucocephalus from

Santa Rosa Island, Okaloosa County, and P. p. rhoadsi from

Archbold's Biological Station, Highlands County.

The mice were kept in cages similar to those described


by Layne (1958); the sides were made of one-quarter inch hardware

cloth and the wood tops were removable. Each cage was placed on

a tray covered with sawdust.

The ambient temperature in the laboratory was 240 20C

most of the time; occasionally it varied as much as 5C. Rela-

tive humidity was measured with a sling psychrometer at different

times during the day and night; it varied from 54 to 84 percent.

Overhead fluorescent lights were automatically turned on at 0630

and off at 2030; no outside light entered the room. Purina

laboratory chow and water were supplied ad lib.

Laboratory-reared mice gradually replaced the wild stock.

Matings of first generation animals were set up with mice from

different parental lines but the same subspecies. Litter size

and data of birth were recorded for each litter born in the

laboratory. After November, 1964, the shelf position of each

cage was maintained constant for 11 months to study possible

synchronization in reproductive activity.

Subspecific crosses

Crosses were made between P. p. subgriseus and the other

three subspecies, P. g. phasma, P. p. rhoadsi, and P. p.

leucocephalus. The F1 from the P. n. subgriseus by P. a. phasma

cross were used to produce an F2 and were backcrossed to P. p.

subgriseus (Table 6).

The differences between the various crosses were signif-

icant in regards to the number of pairs producing young and the

number of young that survived weaning. The interdependence of

the treatments was relatively low (Table 6). The initial

0 0

'0 o
c 0)

-' C)

o 0 1 0
H) gC C

C) H 4-
0) 0

0'H -

00 G B'-:
0 0l H

C 04 -

C i 0

,-' O C'

SO o

0 C) 0 -' 4 C
Qd- U H

'a Co dU-Q
'H -H O)
* H -1 '

0 C ) 0 :p
0 C) 0
CU 'o nk Cn

IH (f1 H C)

4-' 0 O
C*O 6)
I-H 0 )

OH >C).--

'C H
0 0 f
'-H V

C) O C

,0 0 C )
C )rl *
*n *O

o mcS
* -H '
o) 'J .

0 c


CO) C)
o t-i


*H 0) 0
o 3

0 0C

Hl 0

0 C)

Q 0 0

C 0)

0 0


O 0
o d

0C 0 0

0' i 0 H ti)
H- r-H

I c o 0' C N -


X r

x 0
C) -H x
U C C ).
In X 3
U *H Xr

*H ,0 C),

.1 'C 01 0. X V p-
.0 H X Xx

mH C m o N
N H LO r n H-

10 m 0 rl 0)

n 0e m 40 0

H in 041 H H

statistics were used as justification for the more detailed

analysis presented in Table 7.

More pairs of the same subspecies produced young with a

greater chance of surviving weaning than did the pairs formed of

different subspecies (Chi squares = 10.20 and 16.37, df = 1, and

P less than .01, respectively). This was largely due to a de-

crease in reproductive success in crosses between P. p. subgriseus

and P. p. phasma (Chi squares = 5.68 and 32.17, df = 1, and P less

than .05 and .01, respectively) and partially, to those between

P. E. subgriseus and P. p. rhoadsi (Chi squares = 6.96 and .387,

df = 1, and P less than .05 and greater than .90, respectively).

Survival of the young from the latter cross did not show a

significant decrease as it did in the former. There were not

enough crosses involving P. p. leucocephalus to adequately

interpret this subspecies' fertility relationships with P. p.

Subgrfiseus. The F2 from P. p. subgtiseus by P. p. phasma also

showed a breakdown in reproductive success compared to that of

the parental stocks; the decrease in the number of pairs produc-

ing young approached significance (Chi square = 3.47, df = 1,

and P was greater than .05 but less than .10). A significant

decrease was observed in the postnatal survival of the young

(Chi square = 13.73, df = 1, and P less than .01).

The reproductive success of the P. p. subgriseus was

higher for wild animals than for those reared in the laboratory

(Chi square = 5.32 and 10.70, df = 1, and P less than .05 and

.01, respectively). Two factors could have accounted for this

difference, the degree of antagonism between paired mice and

* II
'0 0 -1
C 0

l 0 II


Ct C

0 r

H > r

O n

) OH

0..-. U



.0 0

0 c r
O .0
0 0 -.

0 '0
S*H (,C

i- O

c a
.o C
H 0 .0

ni Q1

O u
*-l 0*

(Q 0
Hri C

u) 1.4

Q (1f C)

04 00
C 0
04H 'H

00 0

0 O
ES 0

0'0 >
0 (

M 0

S <0
3 HO >

*H 0
.' 3

) 01

O. H

2 O+

H HHr H- ` r 4 -l `

r- Hl l H -l

their general level of activity or both. The wild mice were

paired under natural conditions which allowed them to express in-

dividual preferences in selecting a mate. The pair-bonds between

some of the laboratory-reared mice may have been weak because they

were forced to stay together. This effect probably accounted for

at least part of the decrease in reproductive performance.

Laboratory-reared mice were more active than wild mice in

captivity. The subspecific hybrids were extremely agile. When

their cages were cleaned, they frequently jumped out. Following

the birth of a new litter, females normally spend considerable

amounts of time in the nest. The laboratory-reared mice spent

less time in the nest taking care of their young than did the

wild animals, and the same was true for the hybrids in relation

to the other laboratory stocks (Table 8). The time spent caring

for the young decreases as the females spend more time out of

the nest. Neglected young show higher mortality rates during

the postnatal period than do those that are adequately cared for.

The differences between the reproductive success of the various

stocks seem to be correlated with the females' activity pattern;

the more active females give less care to their young.

The importance of heredity is pointed out by the differ-

ences between the subspecific hybrids and their parental stocks

reared in the laboratory. Differences between mice reared in the

laboratory and in the field and between field-reared mice that

have been in the laboratory for varying lengths of time emphasize

the importance of the environment. Wild mice that had lived in

the laboratory for a long time are more active than freshly caught


CN O0 u1 0
1-H H 0

0 *r4

C *i o

a ))

0 c

0) 0 "
C -H 0

H > C

o 0

0 1


0 O 0

,0 x

C) H'0

r .
H *-H

H 0
) -H
*0 H
o C-

(a) H -

.0 H0
o .C +-
01 *8 01

O H0

cC Or
CO 0
(D* r
r-l H


0 t

H 3

0 x

H j




A -I


C'~'0 0 '
m 9 ~o H


0' 00

00 oo
e- i0

O 'H C
C cC

e H

(H ) 0)4
u rC
3 ~ 3 )-
2; 6 H

Inheritance of white cheeks

According to Blair (1944), the inheritance of white cheeks

in the old-field mouse is controlled by a single gene with the

alleles W and w The homozygous recessive type has brown cheeks

(wcwc). The description of each phenotype is based on the color

of the hair directly under the eye. The percentage of white cheek

in the various populations is given in Fig. 9. The Anastasia

Island population of P. p. phasma is 100 percent white-cheeked,

and 8.1 percent of the Ocala population of P. p. subgriseus are

white-cheeked. Several crosses were made between these two sub-

species (Table 9). The lack of a significant difference between

the observed and expected number of progeny with each phenotype

proves again that this trait is based on one gene with two alleles.

Assortative matings

The effects and interactions of natural selection, differ-

ential migration, mutation rate and genetic drift determine gene

frequencies. Assortative mating can also influence gene fre-

quency indirectly by altering the phenotypic composition of a

population. The occurrence of assortative matings in a natural

population would be indicated if the frequencies of the pair

bonds between the different phenotypes differed significantly

from those expected in a panmictic population. If the pheno-

type was controlled by a single gene with two alleles, one domi-

nant and the other recessive, the expected frequencies resulting

from the pairing of two recessive phenotypes, or a dominant and

a recessive one, or two dominant phenotypes should equal the

frequency of the recessive phenotype squared (R2), 2(1-R) (R),
and ( 2 respectively (Li, 1961).
and (1-R) respectively (Li, 1961).

Fig. 9. Geographical variation in the percentage of mice with
white cheeks. Only samples of 15 or more mice are included.

.31---- --- -- - - (.

\ 5.2
.... O

48 23.---
S30 4.80

10.1n ,-

91 120p

87 86




*H 0


0 0 0 0 0'
0) 0 0 OD ON
41 41 6 i! n

a) 4'.2 *d Q) r'I 0 o

CH 0
(01 4 O 0 0 0 10 0
C)" 01 m .
**H O C

o- r) 0
U)C 41 M
0 x0 i
0 4J) )
I .C 1 .0 + 41 f 0

4 4O N < r

O H Oq 0)

Cracl .0
0 0n iO

Q 03

C) ri U)
H0 U0 LX C)
0 -1 H 4 I

a, >, (/) (n-4
S1 41 a)0C
rd Q -> a)

A 0 .* 01 X

0( C), V' ) U)
U) (-4 41Q4
0 0
SX- X X-OX 0 0

o- n o U
H4 O C) u)
4 ( A X H

a)) 60' C)) I
0)0 3 4- 4-c
4) 33 0 z H A

C )0 & 0 H. U1 U4


The "white cheek" trait can be used as a marker to

determine the existence and extent of assortative mating in

populations in which the frequency of the trait is not 0 or 100

percent. The sample size necessary to test the existence of

assortative mating becomes extremely large as the frequency of

the trait approaches either 0 or 100 percent. Only two of the

populations that were sampled met the requirements of adequate

sample size and intermediate phenotypic frequencies. These were

P. p. subgriseus from the Ocala National Forest and P. p. rhoadsi

from the Highlands County, Florida (Table 10). The same pheno-

types paired more often, and unlike phenotypes paired less often

than expected if the populations were panmictic. The differences

were significant for both populations; the mice did not select

their mates at random.

Two factors could account for the assortative mating in

these populations. Mice may select their mates on the basis of

cheek color or they inbreed within family groups. The following

experiment was performed to distinguish between these two alter-

natives. Females with brown cheeks were housed together either

with a brown-cheeked male from the same litter and with a white-

cheeked non-sibling male or with a white-cheeked male litter-

mate and with a brown-cheeked non-sibling male. None of the mice

had bred before, and they were isolated from each other for one

month before being placed together. The white-cheeked non-

sibling males were assumed to be homozygous dominant (WcWc).

Their parents had white cheeks and were descendants of white-

cheeked mice that had produced at least 15 offspring, all of

SH 0

4o I 0
U' 0 1 -r-
D H 0

'0 m )
ttH'0 0 N

v, o 0

'00 (H)
i .0 .H 0
CP C',

,4 .c in
(+n 0

0H1 0
d :J -,4 0

S,-I r
.0 0 0

.H 0 1H
a, o o

4o .O K 0u

0 '4 >4- 0
'44 4 J-' [
O 0 D 0 r- 0
0 0 > 4) U (-0

0 0 a) 0

S.0 ,

Oj 4'01
44 oH Ho
+ ,n 4 S O 4'

44 a' H' 44u *
o4 1 4 o
R 0 ) '4
S0( ;4

H 44 0 4-

04 '4-J *O
0' '0 H
.0 '44 P.0

e v o
c yn
* *^ tl


o -H
d I







. 0

o c

0o 0

0 0
n o)
4 C

0) 0 N
4f c



0 p


'o 10`
10 0)
o4 04

.r( 0



z ca,


a .

n4 4


which were white-cheeked. The white-cheeked litter mates were

heterozygous. The phenotypes of the mice in the first litter

produced by each brown-cheeked female were recorded (Table 11).

The preliminary analysis revealed a significant differ-

ence among the frequencies of the various matings that occurred

(Chi square = 8.63, df = 3, and P less than .05). Two additional

statistical comparisons were made to determine the factor which

contributed most to the difference. Females mated with siblings

significantly more often than expected (Chi square = 7.73,

df = 1, and P less than .05), but they showed no preference for

a particular cheek color (Chi square = .64, df = 1, and P greater

than .50). This supports the hypothesis that assortative mating

in the field populations resulted from inbreeding.

Calculation of inbreeding coefficient

Gene frequencies were calculated for the Ocala popula-

tion. The breeding histories of 38 wild white-cheeked mice from

this locality were extensive enough to determine their genotypes.

The mice were selected at random from the field population and

were crossed to brown-cheeked mice. The occurrence of a single

brown-cheeked offspring proved heterozygosity of the white-

cheeked parent. The smallest number of offspring used to prove

the homozygosity of a white-cheeked mouse was 14. Five of the

white-cheeked mice were homozygous dominant and 33 were hetero-

zygous. The expected values calculated from the observed pheno-

typic frequencies, assuming the population to be panmictic, were

.8 homozygous dominant and 37.2 heterozygous. The differences

between the observed and expected values were significant (Chi

Table 11. The number of times brown-cheeked females mated with
brown-cheeked siblings, white-cheeked non-siblings, white-cheeked
siblings and brown-cheeked non-siblings. The choice always in-
volved either the first two possibilities or the latter two and
no combinations thereof. Expected values were calculated for the
Chi square analysis by assuming the null hypothesis. They had to
be adjusted for the choice involving the heterozygous white-
cheeked sibling because a certain percentage of the litters sired
by this animal would contain all brown-cheeked offspring, and
thus, would be counted for the brown-cheeked non-sibling

Mating Choice of Number of Matings
Brown-Cheeked Females Observed Expected Chi Square

Sibling (wcwc) 22 16.5 1.78

Non-Sibling (WcWc) 11 16.5 2.75

Sub Total 33 33.0 4.53*

Sibling (WcWc) 26 19.4 2.25

Non-Sibling (wcwc) 17 23.6 1.85

Sub Total 43 43.0 4.10*

Grand Total 76 76.0 8.63*

*significant at the .05 level

square = 17.48 with Yate's correction for continuity applied to

the analysis, df = 1, and P less than .01). The conclusion

drawn from these results is that inbreeding occurred in the field

population or that the heterozygotes were at a selective dis-

advantage or both. There is no evidence to support or contradict

the second possibility. The results agree with the earlier con-

clusion about inbreeding in the field population, and they allow

the calculation of F, the inbreeding coefficient (Wright, 1951;

Li, 1961).

The number of homozygous dominant mice (Dol = 5) divided

by the total number of white-cheeked mice in the laboratory (38)

equals the relative frequency of this genotype.

(1) Dol 5 13158
Do, + _- .13158.
Dol + Hol 38

Hol is the number of heterozygotes in the laboratory sample.

This frequency should be the same in the field population. The

actual frequency of the homozygous dominant genotype (Dif) in

the inbreeding field population can be calculated by

(2) Dif = .13158 X NWc

NWc is the observed frequency of the white-cheeked mice in the

field population. For the Ocala population,

(3) NW = .081 (Table 10).

Substituting this value in (2):

(4) Dif = .13158 X .081 = .01068.

(5) Dif + Hif = NWc = .081

Hif is the frequency of heterozygotes in the inbreeding field popu-

lation. Substituting the calculated value of Dif into (5) gives

(6) Hif = .081 .01068 = .07034.

(7) Dif + Hif + Rif = 1 or

(8) Rif = 1 Dif Hif = .9190,

where Rif is the frequency of the homozygous recessive genotype

in the inbreeding field population. Gene frequencies are

calculated by

(9) 2Dif + Hif 7
P 2 .047 and

(10) 2Rif + Hif
q = 2 = 953,

where p and q are the frequencies of the dominant (Wc) and

recessive (wc) genes, respectively. In addition,

(11) p2 = Dofc = (.046)2 = .0021,

(12) q2 = Rofc = (.954)2 = .9101 and

(13) 2pq = Hofc = 2(.046)(.954) = .0878.

Dofc, Rofc, and Hofc are the expected genotypic frequencies

calculated for the field population if it were panmictic (Li,

1961). In the inbreeding field population

(14) Dif = Dofc + Fpq,

(15) Hif Hofc 2Fpq and

(16) Rif= Rofc + Fpq.

Adding (14) and (15) gives

(17) Dif + Hif = Dofc + Hofc Fpq.

This reduces to

(18) .00888
F .04388 .202.

A large number of parent-offspring matings in addition to

many matings between siblings had to occur to account for these

results. The F was very high for a naturally occurring popula-

tion. A large F means that the selective pressures against

certain recessive genes will be much more effective in altering

gene frequencies because the recessive lethals and semi-lethals

will be expressed more often than if F were smaller. This

reduces the amount of hidden genetic variation within the popu-


Howard (1949) estimated that up to 10 percent of the

matings in P. maniculatus involved either a parent and its off-

spring or siblings. Even a figure twice this minimal estimate

could not account for an F = .202. The calculated level of

inbreeding was also much higher than would have been predicted

from the results of the experiments in which the females had a

choice of mating with a sibling or a non-sibling male. One

factor that may help explain this difference was the high rate

of population turnover (p 107 ), Unless a mouse disperses a

long distance, it might have to breed with a related animal or

not breed at all during its short life. The laboratory mice

were also isolated for one month prior to the mating preference

test. Siblings would not ordinarily be separated for long

period in the field if they were going to breed with each other.

The magnitude of the error associated with F is not

known, but a small change in the number of homozygous dominant

mice found in the breeding experiments causes a large fluctuation

in F. For example, seven homozygous dominant mice instead of five

increases F from .202 to .278, which is a 37.7 percent increase.

For this reason, the confidence intervals for F are large.


Greater confidence would require that the genotypes of a larger

number of white-cheeked mice be determined. Even though the

exact value of F may differ from that calculated, it is clear

that a considerable amount of inbreeding occurs in the Ocala

National Forest population and probably in the Highland's County

population. This may be characteristic of the old-field mouse.


Ambient temperature is one of the most important factors

in the environment of mammals. The energy required for the main-

tenance of a constant body temperature is increased at low ambient

temperatures. Many mammals partially avoid this problem by low-

ering their temperature during the period of inactivity (e.g.,

see Morrison, 1962; Morrison and McNab, 1962; Morrison and Ryser,

1962). Greater efficiency in utilizing available energy is im-

portant especially when several closely related forms are sym-


The lack of a significant difference between the tem-

perature of the air in the nest cavities of occupied and un-

occupied burrows (p 22) suggest that P. polionotus reduces its

rate of metabolism and body temperature during the day or that

the soil absorbs heat rapidly. Mice in the laboratory do show

a diel body temperature cycle with a mean maximum value of 38.30C

at 0400 and a mean minimum of 36.20C at 1600"(Smith and Criss,

in press). This small temperature difference does not seem

sufficient to account for the lack of heat accumulation in the

burrow. The burrow temperatures occasionally went as low as

12C. It was first thought that these low temperatures might

induce the mice to go into daily hypothermia, but the regulation

of body temperature in the laboratory was the same at high and


low ambient temperatures (Smith and Criss, in press). The pre-

liminary conclusion is that the mice become torpid in their

burrow but are not doing this in response to low ambient tem-


Torpidity in the field

Almost all of the attempts to capture and record the body

temperature of torpid mice failed for two reasons. First, the

mice can raise their body temperature rapidly when their burrow

is excavated. Secondly, the mice do not normally drop their

body temperature to a level equal to that of their environment.

A dry nest was almost a perfect indicator of mice in the escape

tube. When it was slightly warmer than the surrounding soil,

it was even a better clue. The nest must be slightly warmer

than the air in the burrow to stay dry because the relative

humidity in the nest cavity is close to 100 percent. Nests in

unoccupied burrows were always damp. The mouse was the only

source of heat which could account for this difference.

Torpid mice have been found in the field twice. Sixteen

adult mice, four groups of four each, were captured on January 29,

1966. Their body temperatures, measured with a small animal

probe connected to a Yellow Springs telethermometer, averaged

12.80 .30C, while the temperature of the soil at the depth

of the nest cavity averaged 10.50 .10C. The probe was

lubricated with glycerine and inserted into the anus to a

depth of 2 cm. Shortly after their capture, twelve of the mice

became active and their body temperatures quickly returned to

normal. The four inactive animals were placed singly in jars

with perforated lids in an ice box on top of the ice. At the

end of one hour, two of the mice were jumping at the top of

their jars trying to escape. The body temperatures of the re-

maining two animals were 5.30 and 3.60C. One of them was

left in the ice box for three additional hours. At the end of

this time, it started moving slowly and its body temperature

had gone up to 8.0C. Half an hour later, the animal jumped

around the jar, and its body temperature was 39.20C. All of the

torpid animals were sacrificed one week after their capture.

These results show that the mice can go into torpor in the field,

but they give no clues as to the causes.

Torpor and availability of food

When food is supplied ad lib. in the laboratory, mice

do not drop their body temperature in response to low ambient

temperatures (Smith and Criss, in press). The availability of

food may be the important factor which determines their response

in this situation. This possibility was investigated with 24

wild adult mice from Archer, Florida. Each animal was isolated

from the others. Half of them were fed ad lib. during their

first night in the laboratory; the other half had no food.

Water was supplied ad lib. to both groups. The body temperature

of each animal was measured the following morning. Food was

then given to all of the animals. None of the mice died.

The groups with food and without food had mean body

temperatures of 37.80 1.1C and 29.30 2 2.80C, respectively.

Most of the starved animals had body temperatures very close to

the ambient temperature. While I was taking the body tempera-

tures of the first few animals, some of the remaining mice seemed

disturbed by my presence. This may have caused the difference

between the mean body temperature of the starved animals and

the ambient temperature (24oC). The mice responded to the lack

of food by reducing their body temperature down to or slightly

above the ambient temperature.

Peromyscus eremicus also reduces its temperature in the

absence of food, but dies when their body temperature falls below

160C. MacMillen (1965) suggests that they are not subjected to

temperatures lower than this during their period of inactivity

because they nest at a shallow depth in the soil during the

winter. The sun increases the temperature during the day at

these depths and the mice are active at night. The old-field

mouse is subjected to lower temperatures during the day because

they nest deeper in the ground.

Concentration of respiratory gases and torpor

Smith and Criss (in press) suggested that mice in

closed burrows reduce their metabolic rate to reduce gas ex-

change which might become critical were they to continue

metabolizing at a high rate. Their observations failed to

substantiate this, primarily because the experimental animals

were kept in an open room with good ventilation rather than in

a confined space like a burrow. The normal concentrations of

these gases can be maintained in the nest cavity only by dif-

fusion. Under certain conditions the ratio: of oxygen to carbon

dioxide in the air of the burrow should be decreased due to the

respiration of the mice. Oxygen concentration of the air in

the nest cavities of occupied burrows varied from 15.0 to 20.5

percent and averaged 18.4 percent.

The metabolic rates of 18 wild P. polionotus and 12

P. leucopus were determined at various concentrations of oxygen.

The P. leucopus, which were included for comparison with P.

polionotus, were laboratory-reared descendants of animals that

were originally caught near Chapel Hill, North Carolina. Half

of the P. polionotus and all of the P. leucopus were tested in

a half quart canning jar (about 100 cc) and the rest of the P.

polionotus were tested in a quart canning jar (about 200 cc).

The oxygen electrode and the temperature probe were inserted

through the lid of a canning jar with their sensitive ends in-

side the closed jar. The bottle was submerged in a constant

temperature water bath after placing an animal inside. The mouse

stood on a wire mesh platform two centimeters above the floor of

the jar (Fig. 10). Percent oxygen in the bottle and activity of

the mouse were recorded every five minutes until the animal

stopped breathing or until the oxygen reached 2 percent of the

gas in the bottle. Activity was visually recorded as inactive,

active, or very active.

The P. leucopus were tested at 240 10C. The P.

polionotus were tested at 100 1C and 240 10C. At 240C,

the metabolic rate of P. polionotus was determined with and

without a carbon dioxide absorbent in the jar. Carbon dioxide

was absorbed in the other experiments by pellets of soda lime

in a screen container placed inside the jar. The rectal



OAB i /#N

^ ---- ^1 y /

Fig. 10. Metabolism apparatus.

temperature of each mouse was taken at the end of each recording

session and was also recorded five times at each ambient tempera-

ture for P. polionotus at 14 percent oxygen.

The expected basal rate of metabolism (BMR) was calcu-

lated using the equation BMR = 3.4W-25 (equation derived from

Kleiber, 1961). The metabolic rate was given in cubic centi-

meters of oxygen per gram per hour. W was the weight of the

animal in grams. The percent of the expected BMR of inactive

P. polionotus and P. leucopus at the different oxygen concentra-

tions are given in Figs. 11 and 12. The metabolic rates of active

or of very active animals are excluded from the data presented in

these figures. The only exception to this was that all of the

values for P. polionotus below 9 percent oxygen were included.

There was still considerable variation in the rate of metabolism

and this was probably due to differences in the activity levels

of the various animals which weru considered to be inactive. There

appeared to be no significant differences between the metabolic

rates of the P. polionotus tested in the 100 cc or 200 cc jars,

or at 240 or 10 C, or with or without a carbon dioxide absorbent.

All of the P. leucopus died when the oxygen had reached

4 to 5 percent. The P. polionotus survived in a 2 percent oxygen

atmosphere and appeared normal after this experience. Several

of these animals were saved for breeding stock and have since

reproduced. The body temperatures of the P. leucopus at 4 to 5

percent oxygen and the P. polionotus at 2 percent oxygen approx-

imated the ambient temperature. At 14 percent oxygen, the P.

polionotus had an average body temperature of 26.80C at an

O o*

1oo 4 (-
So 0o 4 -

0 0 (3) 4
-i / * O

So w 0 Co C

0 o > 0 0 4
o a 9 .-

00 00 *V10

ooo o o o X
o a o *; * 5

o o 0 *
0o a0o ** 9 *

0 * I c o-

0o 0 .~ 0 4

S0 0*0
4* W (10

S> O O a o
0 0 M ( 4' a J

0 ". ,. *- .0
l .
o < cO

00 0 0I a
S8 "* L dO>%
> o o

0 I c
0 ** )s

A B'31-' 3 070o
1 04- 01
a * m U

* 0 * ON K
*I, 0 0 0u
a e I -*
* S -* O
o a a a a a0o

* o 0 o

* I JI 2 *-l
* J o > D
in ,
I ____ -- I ---- N- ) 3 r
0 ~ ~ ~ ~ ~ i 0 c.)

0 0 0
O jv J

*u u

C *0

V C **
/ B0
0 n


/ / *

0 co
/ / o
/ / o c O
/ S


. 0,o


U k


i I I I
p 0

co r N

0310-3dX3 %

ambient temperature of 240C and 12.80C at an ambient temperature

of 100C.

Peromyscus leucopus maintained a metabolic rate of about

100 percent of its expected BMR down to approximately 13 percent

oxygen and then was no longer capable of maintaining it. The

steady decline in its rate of metabolism below 13 percent osygen

was correlated with the decreases in oxygen concentration. In

this range, the mice were probably not able to fully load their

hemoglobin with oxygen.

Peromyscus polionotus dropped its rate of metabolism to

about 55 percent of the expected BMR between 18 to 19 percent

oxygen. Its metabolic rate declined to about 30 percent of the

expected BMR at 9 to 10 percent oxygen. At this point, it

either continued to decline or showed an abrupt increase. All

mice showed decreasing metablic rates below 7 percent. Between

4 to 5 percent oxygen there was a sharp decrease in the metabolic

rate. Very active P. polionotus had metabolic rates over 100

percent of the expected BMR between 18 to 10 percent oxygen and

chose between activity and inactivity down to about 7 percent

oxygen. The sharp increase in metabolic rate around 9 to 10

percent oxygen probably represented an alarm reaction, which

could function under adverse conditions by getting the mice to

leave their burrow before they became suffocated. Peromyscus

polionotus efficiently loaded its hemoglobin down to a concen-

tration of 8 percent, which was much lower than the comparable

figure of 13 percent for P. leucopus.

Structural variations in the hemoglobins of the two

species may be responsible for the differences in their oxygen

loading capacities. Foreman (1960 and 1964) has shown that the

electrophoretic patterns of the hemoglobins are exactly the same,

but there are differences in the tryptic peptide components. The

functional significance of these structural variations is unknown.

The differences between the two species can be understood

as adaptive in relation to their respective habitats. The old-

field mouse is semi-fossorial; P. leucopus is partly arboreal

in some parts of its range, including North Carolina, and does

not dig underground burrows (Taylor and McCarley, 1963).

Peromyscus leucopus is not normally exposed to decreased levels

of oxygen. The higher tolerance to anoxia, the alarm reaction,

and the ability to maintain a high metabolic rate at low oxygen

tensions may be interpreted as adaptations of the old-field

mouse to its semi-fossorial existence.

The extent to which mice become hypothermic in the field

is not known, but the average oxygen concentration of the air in

the burrows, 18.4 percent, was low enough to act as a cue for

torpor. Some animals showed reduced metabolic rates in the

laboratory at this concentration. At 14 percent oxygen their

body temperatures were only a few degrees above the ambient

temperature at which mice have been found torpid in the field.

It seems likely that mice undergo daily hypothermia while in

their burrows, and this phenomenon accounts for the lack of a

difference between the air temperature in occupied and un-

occupied burrows. As many as 14 mice have been found in a

single burrow in which the air temperature did not vary sig-

nificantly from that of the air in unoccupied burrows.

Torpid mice can be aroused by local disturbances. The

sand plug in the entrance of the burrows provides an extra

margin of safety by delaying the advance of a potential predator.

The mice can come out of torpor before a very efficient predator,

e.g., myself with a shovel, can dig them out. While the predator

digs down the entrance tube the mice probably elevate their body

temperature and are able to leave their nest via the escape

tube, which cannot be located from outside until the mice break

through the surface.

The lack of a difference between the metabolic rates of

P. polionotus with and without a carbon dioxide absorbent was

unexpected. Instead of breathing more rapidly as the carbon

dioxide concentration increased, most of the mice went to sleep

and continued to breath at the same rate or more slowly. This

is probably a complementary adaptation which goes along with

the reduction in metabolic rate in response to decreased oxygen

tension. If an oxygen gradient was established by a respiring

animal, a carbon dioxide gradient in the reverse direction

should also be established. The higher concentration of carbon

dioxide next to the mouse would normally cause an increase in

the breathing rate, and thus, stop the mouse from becoming torpid.

Increased concentrations of carbon dioxide in the air

that a mammal breathes indirectly change the affinity of hemo-

globin for oxygen by changing the pH of the blood (Bohr effect).

The oxygen equilibrium curve shifts in the direction of higher

partial pressures of oxygen needed to load the hemoglobin upon

addition of acid. The magnitude of the Bohr effect tends to be

greater in small,active mammals than in large,sluggish ones

(Prosser and Brown, 1961). At lower than normal body tempera-

tures, however, the oxygen dissociation curve shifts in a direc-

tion that would partly compensate for the other changes. The

way in which the old-field mouse has solved these problems is

not known, but it may be expected that the molecular structure

of its hemoglobin has been altered and that either the buffering

capacity of the blood is high or the mice are not adversely

affected by fluctuations in pH.

Significance of torpor

Brower and Cade (1966) state that "species of Peromyscus

seem to be characterized physiologically by labile body tempera-

tures." These results support their view. The ability to give

up the regulation of body temperature without suffering permanent

damage is an adaptation of primary importance. It allows a

homeotherm to cut its rate of energy consumption and thus survive

for longer periods of time in a harsh environment. The old-

field mouse may use this adaptation to cut its daily energy con-

sumption in the field.

Peronyscus polionotus from the Ocala National Forest

differs from some of the other species in this genus, in that

it does not use ambient temperature as a cue to reduce its

body temperature (Morrison and Ryser, 1959; Cade, 1964; Smith

and Criss, 1964 and in press; Smith, 1965c; Brower and Cade,

1966). Low ambient t,.peratures about 0C do not occur very

often at this locality.

Mathematical models (e.g., McNab, 1963) which attempt

to describe the daily energy expenditure of a small rodent in

the field must take into account the possibility that its body

temperature may be relatively unstable, especially if it is a

Peromyscus. The magnitude of the error is compounded when the

results of laboratory experiments are extrapolated to the field

situation. For these reasons, the estimate of respiratory

energy flow of P. polionotus in the old-field ecosystem made

by Odum, et al. (1962) is probably too high. Better estimates

will require a knowledge of how frequently and to what extent

the mice forego homeothermy in the field.


Mammals usually have well defined reproductive cycles.

Offspring are produced at times when the probability of their

survival is high. This implies that there is a relationship

between the internal physiological processes and environmental

factors which affect survival. The purpose of this chapter is

to describe certain aspects of this relationship in the old-

field mouse.

Results of field work

The monthly variation in the reproductive characteristics

of the females and males are given in Tables 12 and 13. Litter

size and mean monthly litter size range from 1 to 7 and 0 to

6.2, respectively. The maximum values were greater than the

number of mammae per female (six). Up to 33.3 percent of the

females became pregnant while nursing a previous litter. The

size of the nursing young and the embryos usually indicated

that the pregnancy was the result of a post-partum heat.

The maximum percentage of reproductively active females

was 93.3. This figure was a minimal estimate since any female

that was imperforate and in the early stages of pregnancy would

be excluded from this category. Maturity was determined by the

lack of the juvenile pelage or the presence of embryos. Sub-

adult females cannot be separated from the adults by pelage


0 00
N O rl N

0 t, '0 Un
S N '0 t N

n 11








> H





C a)



C 'H -H
2 3"



4COO r

(0 0D


al 4 a

Z r -4
0 0 H


0 0)

0 3

1 0


O 0

0 OH

&* -


00 0 0 0
'0o o o

C0) Q C

b m

0 0 0 t, 0
0 0 0 0

S \ o0 H-
d d d H

1-1 m

o' i

o 0 o
o a to

0 0 r

o 0 H

0' C' '

C0' cl N i) 0 f ir H H H O
H 4 H r-H r-H Hi H H H

'0 '0 '0 '0 '0
H- C'N C) *< i

o cn m
0 00
o o

0 0 0 ) 0 N1 CO N) 0 0 0

'C U) ,u C) CC)

n n C'

) m (1 0 C 0' cq c- V) C*
H-I CC) ) C c) c) c) C CC C H C C


401 (

4 M


10 u


C ) I' O *


4 J

ri r

00 0
30 0
1 >


] cC1


ID~ Nl
(D 3 )

0 0 H a0
U) Ce C

0 0
o 0

0 Cl) '0

0 CO e0

L- r- N



o o 0 0
.-i i-

0 t 0
o o 0
H* *
O c O

o o
U) .F; a


0 0
o 0

e N (
CCo CC i-
'0 CC

N N t- CC) 0 m C -i H Le) 0 '0 (ON '4

\y 0 \0 \0 U ) U) U) U) Uf) U) U) U )
\ I I I \.0 0 '0 \0 .o 0o 0o '0
I I O 'H- C I I I I I I
-0 0- H H H-4 H C\) m U) t- w 0'

o C '0D o' 0 0 -


0 %0 -. t-' t- t- t, t-. t, 10 t-

S 0 0 N 0 t-0 N O O r-0 C 0

cO L m C i' H N 0n n '-i M 0
0 CO' t- 0 CO C' 0 C N 0

' o o0 co N 0 '0 0 Lf N C' t t tN

Sm mn en m n mn cn It
'0 %0 '0 '0 10 '0 10 '0 '0 '0 D0 '0 '0 '0
I I I I I I I I I I. I I I I
. N t C' O 0 H m I C 0 n Lfn %0
-H H H r-1



*H X

I 3



1 ,H


0 H

I o

'4-4 U
0 -q-4


4 )

0 4


*. X









oA N- i t oi o r- o L o- LA


CA 00 CN C ol %0 O' U)) 0

o t Ln %0 0 0 a, 0 CA 0 Lf

O 4 H t- 0 0 6 LA 4 0 H

0 0 0 10 "0 "0 1 0 1 0 "0 "0 10
o IH 4 C H H H H '

O 0 H C' H CO CA l LA CO LA 0'
o' o ii t o cn ^'i N co o

characteristics, and thus, the percentage of reproductively

mature females would be lowered even more during certain times

of the year.

None of the juvenile males was found with sperm. The

highest percentage of adult males capable of fertilizing a fe-

male was 100. This did not differ much from the maximum per-

centage of reproductively active females, and these peak values

occurred at the same time of the year. The percentage of adult

males with sperm never fell to 0 at any time during the study

as did the percentage of reproductively active females. There

appeared to be a seasonal variation in all of the reproductive

parameters of both sexes with the maximum values occurring

during the winter and the minimum during the summer. The summer

decline may have been due to the sterilizing effect of high

temperatures (Cowles, 1965). Another breeding peak occurred

during the spring each year.

Correlation between production of young, rainfall, and ambient

The monthly variations in litter size, percentage of

reproductively active mature females, amount of rainfall, and

various soil and air temperatures in the field are given in

Fig. 13. Certain aspects of the relationship between rainfall

and ambient temperature have already been pointed out (p 22;

Figs. 7 and 8). A multiple correlation analysis was used to

separate and determine the magnitude of the effects of tempera-

ture and rainfall on reproduction. The dependent variable (P)

was the number of young being produced per 100 mature females

0)4-'C) 4
SC *0 0

0 C'
'O r; ID

R 'H
* 0
. ,C O
H 0

H}) QN

a) CH
Sr4 0 NC O)
,C > 9 a)

H0 ) 0 0 .
>P *4


-)*H cc 0
4C 01 0.

m >3 +1

0- n i (i i


,C 01

m i
80 0

C) C
4) C- 1 '

C '

H 0

0 ~4
CP O )'0

N 0

0 WH

C) O
aR H N

010 C)H

0 0 LO 0 n0 r
o o) ) o' N -


-- 3zIS 83111n
o, n F cj 0

i i- I I- I I- h I I
0 00000000000
0 N3 I3 N 0o %
.- -- S3JVU3N3 3flnJYW dO %

SI ) m I- a q
riO ro N CO -
Oo0, 3nIVy3d '3
I~t Iy
me ememebO~




(Table 12). The independent variables were soil temperature

at the depth of the nest cavity (T) and amount of rainfall

each month (A). The relationship between these three variables

was given by

P = 268.7 1.02T 3.82A.

Rp.TA = .626 (highly significant),

rPT.A = -.581 (significant) and

rPA.T = -.105 (not significant),

where the partial correlation coefficients are rp.TA and rpA.T'

and the multiple correlation coefficient is Rp.TA. The prob-

ability that the amount of variation in the dependent variable

was due to chance rather than to the effects of the independent

variables was less than .01 (F = 7.39 and df = 2/23). The co-

efficient of determination, which is the percent reduction in

the sum of squares of P attributable to the combined effect of

T and A, was 39 percent. This was quite high considering that

the small sample sizes in some months should have increased the

experimental error associated with the sampling technique. Even

though the partial correlation coefficient rPA.T was not signif-

icant, the amount of information given by A about P which was

not given by T was significant (F = 5.68, df = 1/23, and P less

than .01). However, A gave only .39 times as much information

about P as did T. This means that variations in both temperature

and rainfall are important in predicting the number of offspring,

but information concerning temperature is more useful in making

these predictions than information about rainfall.

One obvious variation in reproduction, which was not

revealed by the above analysis, was the consistent spring

breeding peak in the field population (Table 12 and Fig. 13).

Both litter size and percentage of reproductively active females

showed an increase during this season even though ambient tempera-

tures were increasing. Two external factors, a decrease in rain-

fall and'a change of diet, may be associated with this fluctua-


Effect of temperature on the male's reproductive system

Temperature appeared to have an effect upon the repro-

ductive system of the males. More males were reproductively

active and the testes and seminal vesicles were larger during

the winter than in the summer (Table 13). An attempt was made

to quantify this relationship in the laboratory by keeping males

in constant temperature boxes for 60 days at four different

ambient temperatures, 40, 140, 240, 340C with a maximum tempera-

ture fluctuation of t 1C. Food and water were supplied ad lib.

The males were maintained in groups of three to a cage. Their

ages varied from 91 to 108 days at the beginning of the experi-

ment. At this time, their testes were enlarged and in a

scrotal position; the males appeared to be reproductively

active. The temperature that each mouse was exposed to was

randomly selected; nine mice were in each of the four groups.

Another group of the same size and approximately the same age

were selected at random from the laboratory colony; they were

paired with an adult female for at least one month prior to

their death. The 45 males were sacrificed on the same day and

the body weight, size and weight of the testes, and length of

the seminal vesicles were recorded.

Of all the males, only seven, which were paired with

adult females, had their testes in a scrotal position and sperm

in their epididymes at the end of the experiment. Males housed

only with other males and isolated from sensory cues associated

with females became sexually inactive as a result of their

treatment at each ambient temperature. Males paired with females

had significantly larger (42.2 1.1 mm length x width) and

heavier testes (1.30 + .005 g) and longer seminal vesicles

(9.0 1 .21 mm) than the other males (Fig. 14). Either the

presence of a female was necessary for the maintenance of the

male's sexual activity, or the level of interaction between the

males was sufficient to cause the observed effects. In either

case, the pair bond may have functional significance in that

it enables the field population to achieve a high reproductive


Temperature had a significant effect on body weight

(F = 4.52, df = 3/32, and P less than .01), length of the seminal

vesicles (F = 11.4, df = 3/32, and P less than .01), and size

and weight of the testes (F = 6.3 and 16.7, respectively,

df = 3/32, and P less than .01; Fig. 14). Body weight was

lowest at 240C and from this point increased as the temperature

went up or down. This effect of temperature was independent of

food since it was supplied in abundance in the laboratory.

Similar variation may be characteristic of field populations.

Connell (1959) reported that the mean body weight and percentage

Fig. 14. The size of the seminal vesicles and testes, and
weight of the testes and body of males housed with other
males at four different ambient temperatures. There were
nine mice at each temperature. The mean is indicated by the
horizontal line through the dark rectangle which represents two
standard errors. The range is shown by the other horizontal
lines, and the open rectangles indicate one standard deviation
on either side of the mean.



P- -*12
15 --

'/ -.
.08 -

O-, .06 4 l

2 0 -- j

-5 :0


25 40 1

/ "- 3 5 2E

20 25 X

20 2

__15 -j

15 O 5

4 14 24 34

of fat were highest during the summer and winter in a northern

population of old-field mice. In the Ocala National Forest,

acorns and insect parts were usually found during the fall and

spring. If food is locally abundant at approximately the same

time at the northern location, then the mice are apparently

heavier during the time at which food is relatively scarce.

This may be an adaptation which helps the mice to survive during

the lean periods. Survival during short periods when food be-

comes scarce would be assured if the mice could rapidly put on

fat when food was abundant and use this stored energy when

needed. Survival for longer periods could be achieved by

torpor (p 63).

Temperature had a different effect upon the males'

reproductive system. The high values were recorded at 140C,

and decreases were associated with temperature deviations in

either direction from this point, with the only exception being

a slight increase in the length of the seminal vesicles from

240 to 34C. The length of the seminal vesicle is a sensitive

indicator of androgen secretion by the testes in other species

of Peromyscus (Jameson, 1950; Brown, 1964b). Either this was

not the case for P. polionotus or the level of androgen se-

cretion was not related to the weight of the testes in the same

way at the different temperatures. For example, the difference

between the weight of the testes at 240 and 340C was not signifi-

cant, but it was for the length of the seminal vesicles. In

the field, there was no correlation between the size of the

testes and the length of the seminal vesicles, and most of the

males had sperm when their testes were 30 mm2 or more. Shifts

in the age composition of the population may account for some

of the observed seasonal variation. Sexually active subadults

cannot be separated from adults easily; the reported values

are pooled means. Treating the data in this way may fail to

show some of the more significant effects which contribute to

the large variation between the monthly samples, but there is

no way at present to adequately subdivide the data for these

variables, and the sample size per month is not large enough

either. The sensitivity of the seminal vesicles to androgens

is not questioned here, but the relationship between testes

size or weight, androgen secretion, presence of sperm, and the

length of the seminal vesicles cannot be a simple one in this

species. Controlled laboratory experiments must be done before

we can do more than just describe the seasonal variation in these

characters. However, it was clear that the males' reproductive

cycle overlapped that of the females and that most of the time

there were more sexually active males than females in the popu-

lation (Tables 12 and 13).

Minimum critical temperature

Sudden short exposures to 40C stimulated reproduction in

the laboratory (p 95), but prolonged exposures resulted in sig-

nificant decreases in the size and weight of the testes and the

length of the seminal vesicles. Reproduction was highest, and

the testes and seminal vesicles reached their maximum size in

the field, when the soil temperature was slightly below 140C.

The mice never had to endure temperatures below 120C for any

prolonged period in the field. When the mean daily low tempera-

ture dropped below this point the mice could stay in the burrow

or limit the amount of time spent above ground. Temperature is

an important factor regulating above-ground activity as indicated

by decreased trapping success on cold nights (Gentry and Odum,

1957). Prolonged exposure to temperatures below 120 to 140C

has little or no biological meaning for P. p. subgriseus. The

lowest temperature that will stimulate reproduction when exposure

is prolonged must be slightly below 140C.

Breeding performance of Peromyscus polionotus subgriseus in the

The breeding performance of the laboratory colony from

October 1, 1964 to September 31, 1965 is given for each month

in Fig. 15. The mean litter size varied from 3.47 in September

to 4.64 in July. The mean annual value was 4.3 for laboratory-

reared females while the range was one to nine; this was greater

than the range in the field (p 65). The highest percentage of

females giving birth was 76.5 in December with a low value of

26.0 in August. Reproductive performance in the laboratory was

less variable and always greater than that in the field at the

same time. Seasonal variation similar to that shown in the field

occurred in the laboratory despite relatively constant environ-

mental conditions.

Onset of sexual activity

Clark (1938) found that juvenile Peromyscus females went

into their first estrus cycle at as early an age as 23 days, and

that the youngest laboratory-reared female that became pregnant

---- 3ZIS a3111-7



0 0 0 0

CCo ) o 9N A

r cj

was 102 days old. Rand and Host (1942) recorded the earliest

date of reproduction as approximately 48 days in the laboratory

and 40 days in the field. They also found one female 26 days old

with a perforate vagina. In this study, several females that were

less than 26 days old have been found in the laboratory with a

perforate vagina; the youngest was 24 days old. Most of the lab-

oratory-reared females first became perforate between 50 to 60

days if they were paired with a male just after weaning. Some fe-

males never became perforate in their first year of life when kept

in bigger groups, especially in the presence of other adult females.

The earliest date at which conception took place in the laboratory

was 34 days, assuming a gestation period of 23 days (p 83). Other

similar values included 38, 45 and 46-day-old females. All of

these had not yet completed the juvenile molt when they first be-

came pregnant. The data for the field population gave similar

results. The age of each mouse was estimated using the growth

data on the size of the body, tail, hind foot, and ear (Laffo-

day, 1957). Twelve females between 25 to 30 days old had perforate

vaginae when captured. Ten of these were found paired with a sexu-

ally active male, one with a sexually inactive juvenile male and

one in a group of six mice, some of which were in the juvenile

pelage and others were molting for the first time. One of the molt-

ing males had its testes descended but had no sperm. Other similar

values included 39, 41 and 45-day-old females. While some females

bred at an early age, most did not. Over 95 percent of the pregnant

females had completed the post-juvenile molt. The youngest

males that sired litters in the laboratory were 45, 47, 61 and

63 days old.

Post-partum estrus and gestation period

The interval between litters in nonlactating females or

between copulation involving a nonlactating female and the birth

of a new litter was 23 to 24 days; 63 percent on the 23rd day.

Smith (1939), Laffoday (1957), and Williams, et al. (19'65b)

also give 23 to 24 days as the normal gestation period. Non-

lactating females have a shorter gestation period than those

lactating. For the latter, it ranged from 27 to 33 days and

averaged 29.2 f 1.3 days. The comparable figure reported by

Williams, et al. (1965b) is 29.43 t 3.81. Larger litters in

utero and/or nursing seemed associated with longer gestation


Post-partum heat is characteristic of several species

of Peromyscus including P. polionotus (Svihla, 1932; McNair,

1931; Clark, 1936 and 1938; Rand and Host, 1942; Pournelle,

1952; Williams, et al., 1965b). It was of common occurrence in

the laboratory colony (Table 14). Female #58 has given birth

to 26 consecutive litters with a mean inter-litter inverval of

less than 30 days. Her first litter was born when she was 118

days old. Her total production thus far has been 139 young,

and at this time she shows no signs of stopping. Several other

females have also produced over 100 young during the same time.

Post-partum heat was also characteristic of the field population,

and its occurrence accounts for a considerable portion of the

Table 14. Breeding histories of five typical laboratory-reared
females. The size of each litter is indicated in parentheses
after the date of birth

Number of Female

10 15 23 26 58

3- 8-65
4- 4-65
5- 1-65
3- 6-66

4- 1-65
5- 1-65
6- 1-65
7- 1-65
8- 2-65

9- 5-64
10- 6-64
11- 7-64
3- 1-65
6- 7-65
7- 2-65
12- 9-65
1- 6-66
2- 3-66

12- 1-64
5- 6-65
10- 7-65
11- 4-65
1- 7-66
2- 8-66


3- 8-64
4- 4-64
5- 1-64
6- 9-65
7- 6-65
8- 4-65
9- 2-65
10- 1-65

*young were taken from mother at an early age

total reproductive performance at certain times of the year

(Table 12). On two occasions, the laboratory colony increased

more than four times its original size in less than a month.

Synchronization of reproductive activity

The distribution of births in the laboratory was not

random. They occurred in groups with the inter-group time

interval being longer than the average time between individual

litters. To investigate this phenomenon, 60 cages on 12 shelves,

five to a shelf, were maintained in a set position for one year.

The timing and location of each birth was recorded.

There were 363 litters born during this year. Five or

more appeared 31 times in non-overlapping periods of three con-

secutive days. This differs from chance expectation (Chi square

= 336.2, df = 1, and P less than .01). These results support

the preliminary conclusion and give more information about the

probable cause. The daily distribution of litters revealed

that the clumping was not random in respect to the position of

the cages on the shelves (Fig. 16). For example, all of the

females on shelf nine reproduced within a three-day period from

December 11 to 13. Three days is less than the average duration

of the estrus cycle which lasts for about five days (Laffoday,

1957). Since the gestation period was approximately the same

for all females (p 83), they have to go into heat about the same

time if they give birth on the same day. The estrus cycle of

one female influenced the cycle of females in nearby cages. Once

synchronized, the regularity of the post-partum heat and the

uniform length of the gestation period in lactating females

0 0

U) LO tO

to tt
N 0

-) p
rU a 4


-) 0
tI o

0 .0C

~n -' 4



0 4'i S

co 0
------ -' 0

r- 0
.. .14 r

u u

------- (
0 S H

AV) 0 s 1 1
fl /Sd Iif
---- ----C H
0, '" w
-- z; -- Oc "
03 _______ ------ ^1 d
_______ ______ r

AvaO/sd13 in

University of Florida Home Page
© 2004 - 2010 University of Florida George A. Smathers Libraries.
All rights reserved.

Acceptable Use, Copyright, and Disclaimer Statement
Last updated October 10, 2010 - - mvs